PROJECT SUMMARY Pneumonia is a common reason children present to the emergency department (ED) and one of the top three causes for pediatric hospitalization in the United States. Pneumonia also accounts for more days of antibiotic therapy in pediatric hospitals than any other condition. Unfortunately, extensive variation in the delivery of care for children with pneumonia is evident nationally, including differences in antibiotic selection that are often discordant with national pneumonia treatment guidelines, and inconsistent site of care decisions not explained by differences in illness severity. National research priorities for childhood pneumonia emphasize the need for standardized treatment approaches to improve antibiotic stewardship and the creation of objective risk stratification tools to optimize disposition decisions. Our research team recently developed a prognostic tool using prospective data collected from more than 2600 children hospitalized with community-acquired pneumonia that accurately discriminates between children who develop severe in-hospital outcomes and those who do not. Expanding and recalibrating our prognostic tool for use in the ED, where the disease burden and spectrum of illness is greatest, is a critical next step in our research. The ED is also an important setting to test innovative strategies for improving antibiotic prescribing. Few formal antibiotic stewardship programs exist within the ED, and pediatric studies addressing antibiotic overuse for pneumonia in this environment are limited. In addition to our demonstrated expertise in pediatric pneumonia, our team also possesses prior success leveraging health information technology to improve care using clinical decision support that is seamlessly integrated within the electronic health record (EHR). Use of EHR-based antibiotic and prognostic decision support applications in the ED environment are innovative approaches to improve care for children with pneumonia and a central focus of the proposed research. Our multidisciplinary study team, with expertise across the continuum of clinical care as well as in epidemiology, biostatistics, and biomedical informatics, is uniquely suited to conduct these studies. Within two free-standing children's hospitals and one general community hospital, we will test the following hypotheses: 1) Our expanded and recalibrated prognostic tool will accurately discriminate between children with and without high risk for severe outcomes in the ED; 2) EHR-based antibiotic decision support increases guideline-concordant antibiotic use compared with usual care in the ED; and 3) Delivery of EHR-based severity information generated by our prognostic tool improves appropriate site of care disposition in the ED compared to usual care. These hypotheses will be tested within the framework of a prospective, observational cohort and two pragmatic randomized controlled trials. The overarching goal of this research is to demonstrate the potential of risk stratification and high quality EHR-based decision support to improve care delivery and outcomes for children with pneumonia.